wassname/kair_algorithms_draft
1
0 Fork 0
mirror of https://github.com/wassname/kair_algorithms_draft.git synced 2026-10-04 12:40:46 +08:00
Code Issues Packages Projects Releases Wiki Activity

Add DDPGfD, TD3fD and SACfD (#22)

Browse Source 
* Format repository

* Clone files from medipixel repo

* Fix DDPGfDAgent.update_model()

* Fix bug on _initialize()

* Add demo-path parameter and demo data

* Rename init_priority to _max_priority for PER

This makes PER and PERfD consistent.

* Make i_episode attribute of DDPGAgent

* Clone SAC code from medipixel repo

* Fix update_model() for SACfD

* Fix _initialize() for SACfD

* Add is_discrete attribute to AbstractAgent for SACfD

* Add i_episode attribute to SACAgent for SACfD

* Modularize DDPGAgent and SACAgent

* Modify hyperparameters for DDPGfD and SACfD

* Add NStepBuffer

* Add n-step to DDPGfD

* Add n-step to SACfD

* Add TD3fD without n-step

* Attempt to tune hyperparameters

* Remove discrete environment check in SAC

* Implement n-step on TD3fD

* Fix step function of TD3

No done check, and _add_transition_to_memory was not called.

* Fix actor loss calculation for TD3fD

* Attempt to tune hyperparameters

* Print both critic losses

* Fix typo bug

* Attempt to tune hyperparameters

* Fix bug in n-step demo retrieval

* Fix bug in n-step transition addition
This commit is contained in:
Seungjae Ryan Lee authored and GitHub committed 2019-03-14 11:06:54 +09:00
1 parent ee014e5a93
commit ca5c99bc41
17 files changed
+1377 -49
Download Patch File Download Diff File Expand all files Collapse all files

No files matched your search

scripts/examples/lunarlander_continuous_v2/td3.py
+12 -8
View File
@@ -65,25 +65,29 @@ def run(env: gym.Env, args: argparse.Namespace, state_dim: int, action_dim: int)
# create critic1
critic1 = MLP(
input_size=state_dim + action_dim, output_size=1,
hidden_sizes=hidden_sizes_critic
input_size=state_dim + action_dim,
output_size=1,
hidden_sizes=hidden_sizes_critic,
).to(device)
critic1_target = MLP(
input_size=state_dim + action_dim, output_size=1,
hidden_sizes=hidden_sizes_critic
input_size=state_dim + action_dim,
output_size=1,
hidden_sizes=hidden_sizes_critic,
).to(device)
critic1_target.load_state_dict(critic1.state_dict())
# create critic2
critic2 = MLP(
input_size=state_dim + action_dim, output_size=1,
hidden_sizes=hidden_sizes_critic
input_size=state_dim + action_dim,
output_size=1,
hidden_sizes=hidden_sizes_critic,
).to(device)
critic2_target = MLP(
input_size=state_dim + action_dim, output_size=1,
hidden_sizes=hidden_sizes_critic
input_size=state_dim + action_dim,
output_size=1,
hidden_sizes=hidden_sizes_critic,
).to(device)
critic2_target.load_state_dict(critic2.state_dict())